Welcome: Sign up/Sign in page
Main homepage: Insert existing photo/file or use live photo feature
Extracts text from uploaded photo and converts into OpenDyslexic font and formatting.
Customizable user settings bar
Voice features: 3 voice options to read-aloud text
"Summarize with Clarity" feature: Converts text using AI into easy to read bullet points/short sentences
slide 8
slide 1
slide 2
slide 3
slide 4
slide 6
slide 7
slide 5

Inspiration:

Clarity was inspired by someone close to our team who lives with dyslexia. We have seen firsthand how intelligent, capable individuals can struggle not because of their ability, but because of how information is presented. Dense paragraphs, tight spacing, and overwhelming layouts can turn simple reading tasks into exhausting experiences. Through this experience we acknowledge that access to information is a fundamental human right, yet for 780 million people including our friend with dyslexia, that right is obstructed by a "design wall." Most information is structured for neurotypical processing, creating a systemic barrier to education and professional growth. We built Clarity to dismantle this wall, not by simply reading for the user, but by re-engineering information to match how their brains naturally process the world.

What it does:

Our app is an AI-powered accessibility tool designed to help people with dyslexia read physical and digital text more easily. Users can live-scan text through their camera or upload photos (ex. textbooks, worksheets, menus, street signs), and the app transforms the content into a personalized, dyslexia-friendly format.

Technical Architecture:

We architected a multimodal pipeline designed for speed and ethical data handling: -Frontend: Built with React Native (Expo) for a seamless cross-platform mobile experience. -OCR Layer: Utilizes the OCR.space API for accurate raw text extraction from diverse visual environments. -Intelligence Layer: Integrates Gemini 1.5 Flash as a cognitive "simplifier" to restructure dense prose. -Vocalization Engine: Employs the ElevenLabs API for instant, high-quality speech synthesis. -Backend & Security: Powered by Supabase, utilizing Row Level Security (RLS)

Challenges & Engineering Triumphs

Our primary technical hurdle was navigating a complex environment compatibility crisis. We initially hit a "version wall" where our development environment was running Node.js 24, which proved incompatible with our specific build requirements. We had to execute a full downgrade, uninstalling and reinstalling a stable lower version to regain environment control.

This challenge extended to the mobile framework itself; we had to backtrack from the latest Expo 55 release to Expo 54 to maintain library stability for our AI integrations. Navigating these versioning loopholes required deep technical troubleshooting and constant team synchronization. Successfully pivoting through these downgrades was a major win, proving that our team could maintain momentum even when the underlying stack shifted.

Accomplishments & Responsible AI:

We are most proud of our commitment to AI usage in our features. We moved beyond building a standard AI implementation on text to using an OCR to extract text from a live view/photo and also use AI features on it.

Key Insights

The development of Clarity taught us that accessibility is a foundation, not a feature. Designing for the 20% of neurodivergent users inherently creates a more refined, efficient product for the remaining 80%. We learned to leverage Large Language Models (LLMs) for augmentation rather than automation, ensuring AI serves to amplify human agency rather than replace it.

The Road Ahead

AR Overlay: We plan to implement real-time Augmented Reality to project dyslexic-friendly text directly onto physical book pages. ** **Collaborative Ecosystem: Developing a secure platform for students to share "Clarified" study materials. Global Literacy: Expanding our Plain Language models to support non-English speakers facing similar structural barriers worldwide.